#!/usr/bin/env python # # This work is licensed under the GNU GPLv2 or later. # See the COPYING file in the top-level directory. # bug_autorefresh.py: Show what bug_autorefresh is all about, and explain # how to handle the default change via python-bugzilla in 2016 import bugzilla # public test instance of bugzilla.redhat.com. It's okay to make changes URL = "bugzilla.stage.redhat.com" bzapi = bugzilla.Bugzilla(URL) # The Bugzilla.bug_autorefresh setting controls whether bugs will # automatically go out and try to update their cached contents when code # tries to access a bug attribute that isn't already cached. # # Note this is likely only relevant if some part of your code is using # include_fields, or exclude_fields, or you are depending on access # to bugzilla.redhat.com 'extra_fields' type data like 'attachments' # without explicitly asking the API for them. If you aren't using any # of those bits, you can ignore this. # # Though if you aren't using include_fields and you are running regular # queries in a script, check examples/query.py for a simple usecase that # shows how much include_fields usage can speed up your scripts. # The default as of mid 2016 is bug_autorefresh=off, so set it True here # to demonstrate bzapi.bug_autorefresh = True bug = bzapi.getbug(427301, include_fields=["id", "summary"]) # The limited include_fields here means that only "id" and "summary" fields # of the bug are cached in the bug object. What happens when we try to # get component for example? print("Bug component=%s" % bug.component) # Because bug_autorefresh is True, the bug object basically did a # a bug.refresh() for us, grabbed all its data, and now the component field # is there. Let's try it again, but this time without bug_autorefresh bzapi.bug_autorefresh = False bug = bzapi.getbug(427301, include_fields=["id", "summary"]) try: print("Shouldn't see this! bug component=%s" % bug.component) except AttributeError: print("With bug_autorefresh=False, we received AttributeError as expected") # Why does this matter? Some scripts are implicitly depending on this # auto-refresh behavior, because their include_fields specification doesn't # cover all attributes they actually use. Your script will work, sure, but # it's likely doing many more API calls than needed, possibly 1 per bug. # So if after upgrading python-bugzilla you start hitting issues, the # recommendation is to fix your include_fields.